Validation of the German Version of the Brief Fatigue Inventory

Vol. 25 No. 5 May 2003 Journal of Pain and Symptom Management 449 Original Article Validation of the German Version of the Brief Fatigue Inventory Lukas Radbruch, MD, Rainer Sabatowski, MD, Frank Elsner, MD, Jan Everts, Tito Mendoza, PhD, and Charles Cleeland, PhD Department of Anesthesiology (R.S., F.E., J.E.), University of Cologne, Cologne, Germany; and Pain Research Group (T.M., C.C.), University of Texas, MD Anderson Cancer Center, Houston, Texas, USA; Department of Palliative Medicine (L.R.), University of Aachen, Aachen, Germany Abstract Sedation and tiredness are among the most frequent symptoms among cancer patients. A detailed assessment of these symptoms is necessary to evaluate therapeutic effects, such as the use of methylphenidate or comparison of different opioids. The Brief Fatigue Inventory (BFI) has been validated as a short and comprehensive instrument to assess severity of fatigue and fatigue-related impairment in cancer patients. We validated the German version of the BFI in patients with chronic cancer-related and noncancer-related pain treated in a tertiary pain center. Patients treated in the Pain Clinic of the Department of Anesthesiology completed the BFI, the minimal documentation system (MIDOS) and the short form SF-36 quality-of-life questionnaire (SF-36). Test-retest reliability was assessed with a second BFI immediately after the consultation and in a subgroup of patients after 3 to 7 days. Nineteen percent of the 117 patients were treated for cancer-related pain (C); the other patients suffered from chronic severe pain of nonmalignant origin (NC). Patients reported mean values for average fatigue of 3.9 (C) and 4.9 (NC), and for worst fatigue of 5.5 (C) and 6.2 (NC). The mean score of the 6 impairment items was 4.3 in both groups. Factor analysis led to a solution with one common factor for all nine items. Fatigue on the BFI correlated highly with feeling tired in the SF-36 and with sedation in MIDOS, and less with being worn out in SF-36 and weakness in MIDOS. Internal consistency was high, as was test-retest reliability, with a correlation of the intensity, mean scores of 0.93 and the impairment mean scores of 0.87. In conclusion, we found the German version of the BFI to be reliable and valid for cancer and noncancer patients. Minor differences were seen in the validation compared to the original version. J Pain Symptom Manage 2003;25:449 458. 2003 U.S. Cancer Pain Relief Committee. Published by Elsevier. All right reserved. Key Words Fatigue, assessment, validation, cancer pain, noncancer pain Address reprint requests to: Lukas Radbruch, MD, Department of Palliative Medicine, University of Aachen, 52074 Aachen, Germany. Accepted for publication: August 5, 2002. Introduction Fatigue has been found to be the most frequent symptom in cancer patients and the most frequent and impairing side effect of cancer therapies. It occurs in a very high proportion of patients undergoing chemotherapy or radiotherapy. Pathophysiological mechanisms that may be involved include anemia, reduction of 2003 U.S. Cancer Pain Relief Committee 0885-3924/03/$ see front matter Published by Elsevier. All rights reserved. doi:10.1016/s0885-3924(03)00073-3

450 Radbruch et al. Vol. 25 No. 5 May 2003 muscle cell tonus from vagal stimulation, central serotonin dysregulation, and generalized lack of high energy phosphatases such as adenosine-3- phosphate. 1 Any additional factor such as fever, infection or psychosocial stress may further drain the reduced ressources of energy and increase the feeling of fatigue and lack of energy. Fatigue is one of the main symptoms of major depression. 2 Although there is no generally accepted definition of fatigue, it is recognized as a multidimensional construct involving physical exhaustion, mental tiredness and a lack of energy. 3, 4 Several instruments have been constructed for the multidimensional assessment of fatigue. Some are very long, such as the Multidimensional Fatigue Inventory (with 58 items) 5 or the Functional Assessment of Cancer Therapy Fatigue (FACT-F) (with 47 items), 6 or they do not adequately cover the different dimensions of fatigue. The Piper Fatigue Inventory covers the various dimensions but even the revised version remains relatively long at 22 items. 7 The Brief Fatigue Inventory (BFI) was constructed to assess severity and impairment from fatigue in nine questions. The English version has been validated in cancer patients treated in oncologic outpatient and inpatient samples and in healthy controls. 8 The BFI was found to be a valid and reliable instrument tapping a single dimension described as severity of fatigue. It correlated highly with other fatigue assessment tools. In a clinical trial using the BFI, Wang et al. found pain severity to be the only significant predictor of severe fatigue in cancer patients. 9 However, no comparative data have been described between patients with cancer-related and noncancer-related pain. Fatigue is not restricted to cancer and its therapies. A prevalence of up to 28% has been found in unselected patient groups treated by general practitioners. 10, 11 Comparing cancer patients with healthy controls, Mendoza et al. found similar rating in several fatigue instruments, with controls rating fatigue considerably less consistently in all instruments. 8 Evaluating fatigue in disease-free cancer patients, Servaes et al. found a subgroup of patients with severe fatigue comparable to that of patients with chronic fatigue syndrome. 12 In the German-speaking countries, the working group of Glaus et al. has provided information on fatigue in several trials. In a qualitative study they investigated cognitive concepts related to fatigue and found similar concepts in cancer patients and healthy controls, with physical, affective and cognitive issues in both groups. 3 However, healthy controls described fatigue rather as a pleasant acute and normal phenomenon, whereas cancer patients reported it as an unusual tiredness with weakness, decreased physical performance and an unusual need for rest. Validating a visual analogue scale for fatigue, Glaus et al. found a higher mean intensity of fatigue in cancer patients compared to noncancer patients with chronic inflammatory gastrointestinal disease and healthy controls. Whereas healthy controls started the day without tiredness and became fatigued in the evening, noncancer and cancer patients started in the morning with higher levels of fatigue, which increased throughout the day in the cancer patients, but not in noncancer patients. 13 Apart from this visual analogue scale and the 20-item Fatigue Assessment Questionnaire available only in German, 4 no fatigue instrument has been validated in the German language up to now. In this study we validated the German version of the BFI (Appendix 1) in patients with cancer pain and noncancer-related pain treated in a tertiary pain center. Methods The BFI consists of an initial question asking the patient whether he feels more than usual fatigued or tired; three questions asking for the intensity of fatigue on the average, at its worst and right now; and six questions asking for the impairment of general activity, mood, walking ability, normal work, relations with others and enjoyment of life. Eleven-step numerical rating scales are used for the items on severity and impairment, with higher scores signifying higher intensity and impairment. The original version was translated and back-translated by a native speaking physician who did not know the original version. Discrepancies between the back-translation and the original version were then discussed and the German version modified accordingly (Appendix 1). The questionnaire for this study was constructed containing the BFI, the minimal documentation system for palliative care patients (MIDOS) 14 and the SF-36 quality-of-life questionnaire (SF-36) 15 in the first part, and a retest of the BFI as well as open questions about the

Vol. 25 No. 5 May 2003 Validation of the German Version of the Brief Fatigue Inventory 451 time needed for completion of the questionnaire, difficult items and patient satisfaction with the BFI in the second part. Physicians scored performance status using the Eastern Cooperative Oncology Group (ECOG) scale, which ranges from 0 no impairment, fully capable of work, to 4 bedridden, not able to care for oneself. 16,17 MIDOS consists of two numerical rating scales for average pain and worst pain intensity (0 no pain, 10 as bad as could be), categorical scales for sedation, nausea, constipation, dyspnea, weakness and anxiety (none, slight, moderate, severe) and a categorical scale for well-being (very bad, bad, average, good, very good). Patients were asked to rate these items for the day of assessment and the day before. Sum scores may be calculated from the two pain items and from the symptom and well-being items. Patients waiting for their appointment in the Pain Clinic of the Department of Anesthesiology were approached and asked to participate in the study. Newly-admitted patients were included as well as patients who were already treated in the Pain Clinic and waiting for follow-up controls. After written informed consent patients completed the first part of the questionnaire in the waiting area. After the appointment, which took on average 30 minutes, the physician completed the performance status with the ECOG score, and the patient completed Part 2 of the questionnaire. Patients were asked about difficulties and assistance needed for completion of the questionnaire to evaluate the comprehensibility. The time needed for completion was used to describe patient burden. Patients scheduled for follow-up appointments after 3 to 7 days were asked to complete the BFI a third time to obtain information about long-term test-retest stability. Patients were included consecutively and reasons for non-inclusion were documented if patients were not included. It was estimated that factor analysis of the BFI would need approximately 90 patients (10 patients per item). With an estimated drop-out rate of 25%, inclusion of 120 patients was targeted. Data from the questionnaires were transferred to a computerized data base. Statistical evaluation was performed with the statistical package SPSS. Factor analysis was conducted using principal axis factoring to extract factors and oblimin rotation to allow for dependency between the identified factors. T-tests were used for comparisons of cancer and noncancer patients for items with numerical scales and mean scores of intensity and impairment. Mann- Whitney U-test was used for categorical items. Spearman correlation coefficient was calculated for correlation of categorical items. Internal consistency was tested with Cronbach alpha and alpha if item deleted, test-retest reliability was evaluated with Pearson correlation coefficient. Categorization of the worst fatigue intensity was calculated with multivariate analysis as described by Mendoza et al. 8 Worst fatigue was recoded with different boundary models for mild, moderate and severe fatigue and F-values for Pillai s trace, Wilk s lambda and Hotelling s trace were calculated. Larger F-values in these tests should be associated with cut points that maximally discriminate fatigue severity. The study was approved by the Ethical Committee of the University of Cologne. Results Problems arose from the translation of fatigue, as there is no close German translation for this term, and though the word fatigue has been used in chemotherapy trials by German authors, it is not commonly used by physicians or lay people. Options for translation may not always cover the different dimensions of fatigue and care has to be taken that the translated questionnaire covers the same dimensions as the original versions. Glaus et al. stated that for German-speaking patients fatigue would be translated with Ermüdung or Ermüdbarkeit, whereas Müdigkeit would rather be equivalent with tiredness. In a qualitative study, her working group has shown that German-speaking patients associated physical, affective and cognitive areas with fatigue. 3 In our translation, we chose Ermüdung as the equivalent for fatigue, Müdigkeit for tiredness and Mattigkeit for weariness. One hundred and seventeen patients were included in the study (74 women, 43 men). Twenty-two patients suffered from cancer pain and 95 from noncancer-related pain. Noncancer pain syndromes were low back pain (39%) and neuropathic pain (34%) mostly (headache 13%, musculoskeletal pain 7%, ischemic pain 3%, others 4%). Patient ages ranged from 21

452 Radbruch et al. Vol. 25 No. 5 May 2003 Table 1 Brief Fatigue Inventory a Cancer Pain Noncancer Pain All Patients Fatigue right now 4.5 2.2 4.3 2.6 4.4 2.6 Usual fatigue 3.9 2.1 4.9 2.5 4.7 2.5 Worst fatigue 5.5 2.6 6.2 2.3 6.1 2.4 Activity 4.4 2.6 5.9 2.6 4.9 2.6 Mood 4.4 2.5 4.3 2.6 4.3 2.6 Walking 4.1 3.2 4.0 3.3 4.0 3.3 Working 5.2 2.9 5.2 3.0 5.7 3.0 Relation to others 3.3 2.8 3.3 2.7 3.3 2.7 Enjoyment of life 4.3 3.0 3.9 2.8 4.0 2.8 BFI intensity means score 4.6 2.0 5.1 2.3 5.0 2.2 BFI impairment mean score 4.3 2.4 4.3 2.3 4.3 2.3 BFI total mean score 4.4 2.1 4.6 2.1 4.5 2.1 a Mean standard deviation, numerical rating scales 0 none/does not interfere, 10 as bad as you can imagine/completely interferes Table 2 Minimal Documentation System (MIDOS) a Cancer Pain Noncancer Pain All Patients Average pain 4.2 3.0 5.5 2.5 5.3 2.6 Worst pain 4.8 3.0 5.5 2.9 5.4 2.9 Sedation 2.7 0.8 2.7 1.0 2.7 0.9 Nausea 1.3 0.6 1.4 0.6 1.4 0.6 Constipation 1.9 1.1 1.8 1.0 1.8 1.0 Dyspnea 2.0 1.1 1.4 0.8 1.5 0.9 Weakness 2.7 1.0 2.1 0.9 2.2 0.9 Anxiety 1.8 1.0 1.5 0.8 1.6 0.8 Well-being 2.9 0.9 2.9 0.8 2.9 0.8 a Mean standard deviation, pain items: numerical rating scale (0 10), symptoms: categorical scales (0 3), well-being: categorical scale (0 4) to 84 years. Cancer patients were slightly older (mean age 59.5 12.6) than noncancer patients (55.1 14.4). Mean ECOG performance scores were 1.2 0.7 (cancer 1.6 0.9, noncancer 1.1 0.7). Half of the cancer patients and 38% of the noncancer patients were treated with Step-3 opioids and 18% (cancer) and 23% (noncancer) with Step-2 opioids. Antidepressants were given to 9% (cancer) and 10% (noncancer) of the patients, anticonvulsants to 14% (cancer) and 36% (noncancer). Patients with noncancer-related pain scored higher values for usual and worst fatigue and for fatigue-related impairment of activity in the BFI, and lower scores for impairment of enjoyment of life, though these differences did not reach statistical significance (Table 1). Fatigue scores of 7 or more were reached by 22% of the noncancer patients for fatigue right now, 30% for usual fatigue and 45% for worst fatigue, compared to 19%, 10% and 35% of the cancer patients. Noncancer patients also reported higher pain scores in MIDOS, but less dyspnea, weakness and anxiety. Only the differences in dyspnea (0.018) and weakness (0.011) reached significance in the Mann-Whitney U- test (Table 2). The physical component score of the SF-36 was 29.4 8.6 and the mental component score 42.4 11.9, with no significant differences between cancer and noncancer patients. Content validity was predetermined from the English version. Construct validity was confirmed by factor analysis, showing two factors with eigenvalues over 1. The three items on fatigue intensity loaded on the first factor, and the impairment items on the second factor. However, the solution with two factors was not interpretable, as factor loadings for two items namely impairment of mood and activity had high values on both factors. Model fit for a solution with one factor (Table 3) was assessed with calculation of the residual correlations (difference between actual correlations and reproduced correlations). A solution is considered adequate if the standard deviation of these residuals is less Table 3 Factor Analysis, Principal Axis Solution Factor Fatigue right now 0.780 Usual fatigue 0.812 Worst fatigue 0.737 Activity 0.888 Mood 0.831 Walking 0.611 Working 0.746 Relation to others 0.682 Enjoyment of life 0.767

Vol. 25 No. 5 May 2003 Validation of the German Version of the Brief Fatigue Inventory 453 Fig. 1. BFI mean scores and the item tired from the SF-36 quality-of-life questionnaire. than or approximately equal to the standard error of a correlation coefficient. As the standard deviation of the 36 residual correlations (0.09) was equal to the standard error of a correlation (reciprocal of the square root of the sample size 1/sqrt (117) 0.09), the onefactor solution was deemed acceptable. The values of the factor analysis were only marginally different for the subgroups cancer and noncancer patients. Mean scores of the intensity items and the impairment items correlated highly (Pearson correlation coefficient 0.70). Cut points for severe intensity on the worst fatigue scale were estimated with multivariate analysis. Cutting between 7 and 8 resulted in larger F-values than a cut point between 6 and 7. Multivariate analysis did not differentiate clearly between the cut points 3/4 and 4/5 for mild and moderate fatigue. For evaluation of criterion validity the mean scores of the BFI were compared to MIDOS, SF-36 and to the ECOG performance status. BFI mean scores showed consistency with the items tired (Spearman correlation coefficient with intensity: 0.67, correlation with impair- Fig. 2. BFI mean scores and the item worn out from the SF-36 quality-of-life questionnaire.

454 Radbruch et al. Vol. 25 No. 5 May 2003 Fig. 3. BFI mean scores and the item sedation from MIDOS. ment: 0.58) and worn-out ( 0.52 and 0.51) in the SF-36 and with the items sedation (0.76 and 0.61) and weakness (0.46 and 0.52) in MIDOS (Fig. 1 Fig. 4). The mean scores showed higher correlation with the mental component score ( 0.43 and 0.57) than the physical component score ( 0.20 and 0.37) of the SF-36. However, with decreasing rating of the performance status both BFI mean scores also did increase (Table 4). Neither sex nor medication with opioids, antidepressants or anticonvulsants led to significant differences in the BFI mean scores. Patients who affirmed the first question of the BFI (other than everyday fatigue) reached significant higher levels in the mean scores (intensity 5.6 1.9, impairment 4.6 2.2) than those who negated it (intensity 2.8 2.0, impairment 2.9 2.4). Reliability was good with high Cronbach s alpha for the BFI and high alpha if item deleted for all nine items (Table 5). Test-retest reliability was high also. Correlation of the intensity mean score (Pearson correlation coefficient) was 0.93 between first and second BFI and 0.75 between first and third BFI, and correlation of Fig. 4. BFI mean scores and the item weakness from MIDOS.

Vol. 25 No. 5 May 2003 Validation of the German Version of the Brief Fatigue Inventory 455 Table 4 Brief Fatigue Inventory Mean Scores and Performance Status (mean standard deviation) ECOG Patients Intensity Impairment ECOG 0 16 4.2 2.0 2.3 1.5 ECOG 1 70 5.2 2.3 4.2 2.1 ECOG 2 20 5.1 2.2 5.6 2.0 ECOG 3 4 6.3 2.2 7.8 2.2 ECOG 4 1 2.3 1.5 impairment mean score was 0.87 between first and second BFI and 0.81 between first and third BFI. Correlation of the total BFI mean scores was 0.91 between first and second BFI and 0.79 between first and third BFI. Mean time needed for completion of the first part of the questionnaire with BFI, MIDOS and SF-36 was 12 minutes, and for the second part with the retest of the BFI and comments 5 minutes. Only 4 patients took more than 10 minutes for this part. Eight patients reported that they needed assistance for the completion of the questionnaire. Sixteen patients reported problems in understanding with single items or the complete questionnaire, though this was related to the SF-36 in most cases, and only in two cases with the BFI. Many patients offered vivid comments on their fatigue. These comments often focused on cognitive impairment, complicating activities such as the reading or driving a car. Patients complained about persistent fatigue, often with affective descriptions such as listless, depressing, irksome or paralyzing. Physical weakness was described less often. Some patients stated that fatigue reduced their quality of life, others complained about the activities that they could not Table 5 Reliability with Cronbach Alpha and Alpha If Item Deleted First BFI Second BFI Third BFI Patients 117 101 16 Alpha 0.922 0.924 0.949 Alpha if item deleted Fatigue right now 0.913 0.915 0.942 Usual fatigue 0.911 0.919 0.937 Worst fatigue 0.915 0.920 0.946 Activity 0.905 0.907 0.938 Mood 0.910 0.909 0.938 Walking 0.924 0.925 0.952 Working 0.913 0.915 0.937 Relation to others 0.916 0.915 0.943 Enjoyment of life 0.912 0.912 0.949 perform because of fatigue. Only few patients remarked that they had gotten used to their fatigue, but only two patients found it pleasant. Only two comments on what the patients had missed in the questionnaire were related to fatigue: What does the patient think why he is fatigued? and What is the impact of fatigue on family life?. Discussion Fatigue and tiredness are among the most frequent symptoms found in cancer patients. 18 Fatigue (84% of the patients), weakness (66%) and lack of energy (61%) were three of the five most frequent symptoms in a study with 1000 patients in an American palliative care program. 19 The impact of these symptoms on the patients can be very high, as there are only few therapeutic options for the relief of fatigue. 20 Fatigue often is associated with the opioid therapy by patients and caregivers, though the malignant disease itself as well as many other factors such as electrolyte imbalances, sleep disturbances or other adjuvant medications can contribute significantly to the genesis of fatigue. The English term fatigue includes concepts such as tiredness, drowsiness and weariness on to exhaustion, though no generally accepted definition of fatigue has been found yet. 21 The translation into German is not without problems, as the term fatigue is not commonly used in Germany. Ermüdung or Ermüdbarkeit seem to be the best options for translation of fatigue. 3 Tiredness and weariness are associated with physical, affective and cognitive themes by German-speaking patients. 3 Asking for the intensity of fatigue in a single question, as included in symptom checklists such as MIDOS 14 or quality of life questionnaires, therefore can yield different responses depending on the prevailing concept of fatigue in the mind of the patient or the caregiver. On the other hand a detailed assessment of fatigue is necessary to evaluate the effectivity of different therapeutic regimens or to compare the side effects of different drugs such as opioids or antidepressants. The problems with the translation as well as cultural differences in the concepts associated with fatigue may have led to differences in the validation of the German as compared to the original version. The factor solution for the German version produced two factors with eigen-

456 Radbruch et al. Vol. 25 No. 5 May 2003 values over one, whereas in the original version all items loaded on a single factor. 8 However, this solution with two factors was not interpretable and model fit for a one-factor solution was good. Therefore using a total BFI score with the mean of all nine items should be appropriate for the German version as well. The increase in fatigue severity with decreasing performance status was less in our study than in the English version, where mean scores increased from 2.3 (ECOG 0) to 4.0 (ECOG 1) and 6.0 (ECOG 2). Cut-off point for severe fatigue on the worst fatigue subscale also differed in the two versions, as the best cut-off point was described between 6 and 7 in the English version and between 7 and 8 in our validation. Part of these differences may also have been due to different patient populations, as patients from oncologic departments were studied in the English version, and this patient group may not be comparable to a group treated in a pain clinic. Patients in our study clearly related fatigue more with cognitive and affective areas than with physical exhaustion. Correlation of the BFI with sedation and feeling tired was higher than with weakness and being worn out, and with the mental component score of the SF-36 higher than with the physical component score. Comments of the patients also pointed in this direction, with more comments focussing on cognitive impairment, sedation and the affective burden of fatigue than on physical weakness. These comments support our translation of the BFI, as the German equivalents of fatigue in the BFI seem to have assessed the predominant areas that the patients have described. Fatigue scores were high in our study, and pain and fatigue intensity were even higher in noncancer patients than in cancer patients, though cancer patients reported more weakness and dyspnea. Wang et al. found fatigue intensities similar to the cancer patients in our study using the BFI in a sample of patients undergoing radiotherapy for rectal cancer. 9 Patients with moderate or severe pain reported mean scores of 6.4 for worst fatigue, though mean scores were only 3.2 for the whole group of patients including those with less pain. Similar to our study, the majority of the patients had high performance scores (ECOG 0 1). Pain intensity was the only significant predictor for fatigue severity in their study, and high pain intensities in our patients may have been related to the high fatigue scores in our patients. Glaus et al. found a higher mean intensity of fatigue in noncancer patients with chronic inflammatory gastrointestinal disease compared to healthy controls, though noncancer patients reported less fatigue than cancer patients. 13 However, the primary aim of this study was the evaluation of the German version of the BFI, and the small number of patients in the cancer group prevented detailed analysis of differences. MIDOS has been validated in palliative care patients with cancer, but has not been used in noncancer patients before. The only significant difference between cancer and noncancer patients was less weakness and dyspnea in noncancer patients. However, as MIDOS has not been validated for noncancer patients, differences between cancer and noncancer patients in the concepts related to the symptoms assessed might have biased the criterion evaluation. Most noncancer patients, as well as the cancer patients, were treated with sedative drugs, and this may have added to fatigue, though the medication was not related to significant differences in our study. Patients with chronic pain may develop certain characteristics that could influence the response to distress, others may develop chronic pain syndromes because of such personal characteristics. These traits could have influenced the patients perception of fatigue and this might have influenced validity and reliability of the data. Psychological assessment and qualitative studies as well as the comparison of different groups of noncancer patients would be required to rule out bias from this source. In conclusion, we found the German version of the Brief Fatigue Inventory to be reliable and valid, with a low burden on the patients. Minor differences were seen in the validation compared to the original version. Patients with noncancer-related pain showed amazingly high scores of fatigue. More research is needed to investigate the high level of fatigue in these patients, which may be related to a higher general level of complaints in these patients or may be part of the pain-related impairment. Advantages of the BFI are its brevity and the clear cut-off points for severe fatigue, marking it as a useful screening and monitoring instrument even for patients with reduced perfor-

Vol. 25 No. 5 May 2003 Validation of the German Version of the Brief Fatigue Inventory 457 mance status. Research is needed with the BFI in palliative care patients with far advanced incurable disease to get more information on the differences between cultures and in different patient populations. References 1. Morrow GR, Andrews PL, Hickok JT, et al. Fatigue associated with cancer and its treatment. Supportive Care in Cancer 2001; Supp Care Cancer 2002;10:389 398. 2. Visser MR, Smets EM. Fatigue, depression and quality of life in cancer patients: how are they related? Support Care Cancer 1998;6:101 108. 3. Glaus A, Crow R, Hammond S. A qualitative study to explore the concept of fatigue/tiredness in cancer patients and in healthy individuals. Support Care Cancer 1996;4:82 96. 4. Glaus A. Fatigue in patients with cancer. Analysis and assessment. Recent Results Cancer Res 1998; 145:I-XI, 1 172. 5. Smets EM, Garssen B, Bonke B, De Haes JC. The Multidimensional Fatigue Inventory (MFI) psychometric qualities of an instrument to assess fatigue. J Psychosom Res 1995;39:315 325. 6. Yellen SB, Cella DF, Webster K, et al. Measuring fatigue and other anemia-related symptoms with the Functional Assessment of Cancer Therapy (FACT) measurement system. J Pain Symptom Manage 1997; 13:63 74. 7. Piper BF. Piper fatigue scale available for clinical testing. Oncol Nurs Forum 1990;17:661 662. 8. Mendoza TR, Wang XS, Cleeland CS, et al. The rapid assessment of fatigue severity in cancer patients: use of the Brief Fatigue Inventory. Cancer 1999;85:1186 1196. 9. Wang XS, Janjan NA, Guo H, et al. Fatigue during preoperative chemoradiation for resectable rectal cancer. Cancer 2001;92:1725 1732. 10. Kroenke K, Wood DR, Mangelsdorff AD, Meier NJ, Powell JB. Chronic fatigue in primary care: prevalence patient characteristics and outcome. JAMA 1988;260:929 934. 11. David A, Pelosi A, McDonald E, et al. Tired, weak, or in need of rest: fatigue among general practice attenders. BMJ 1990;301:1199 1202. 12. Servaes P, van der Werf S, Prins J, et al. Fatigue in disease-free cancer patients compared with fatigue in patients with chronic fatigue syndrome. Support Care Cancer 2000;9:11 17. 13. Glaus A. Assessment of fatigue in cancer and noncancer patients and in healthy individuals. Support Care Cancer 1993;1:305 315. 14. Radbruch L, Sabatowski R, Loick G, et al. MIDOS Validierung eines minimalen Dokumentationssystems für die Palliativmedizin. Schmerz 2000;14:231 239. 15. Bullinger M, Kirchberger I, Ware J. The German SF-36 Health Survey. Zeitschrift für Gesundheitswissenschaften 1995;3:21 36. 16. Buccheri G, Ferrigno D, Tamburini M. Karnofsky and ECOG performance status scoring in lung cancer: a prospective, longitudinal study of 536 patients from a single institution. Eur J Cancer 1996; 32A:1135 1141. 17. Conill C, Verger E, Salamero M. Performance status assessment in cancer patients. Cancer 1990; 65:1864 1866. 18. Grond S, Zech D, Diefenbach C, Bischoff A. Prevalence and pattern of symptoms in patients with cancer pain: a prospective evaluation of 1635 cancer patients referred to a pain clinic. J Pain Symptom. Manage 1994;9:372 382. 19. Walsh D, Donnelly S, Rybicki L. The symptoms of advanced cancer: relationship to age, gender, and performance status in 1,000 patients. Support Care Cancer 2000;8:175 179. 20. Glaus A. Fatigue and cachexia in cancer patients [editorial]. Support Care Cancer 1998;6:77 78. 21. Krishnasamy M. Fatigue in advanced cancer: meaning before measurement. In: Field D, Clark D, Corner J, Davis C, eds. Researching Palliative Care. Buckingham: Open University Press. 2000: 81 97.

458 Radbruch et al. Vol. 25 No. 5 May 2003 Appendix 1 German Version of the Brief Fatigue Inventory